I need to check the spec more closely and consult with the tck folks as
well.
Doug, others, any thoughts on this?
Thanks,
Kathy
Kohsuke Kawaguchi wrote:
> kathy walsh wrote:
>
>> So given the current programming model in JAX-WS rearch,
>> how might you suggest we fix this problem? I would like to get this
>> bug fix
>> in tonight's build without causing side effects. Your quick response
>> would be
>> appreciated.
>
>
> If it's not strictly a requirement to preserve the original behavior,
> I think the right thing to do is to capture such information into the
> exception being thrown. That is, if the HTTP transport decided to
> throw an exception because of 404, then the exception should indicate
> so. We could define a custom exception type that carries related
> information, such as the URL, the HTTP status code, or even the whole
> HTTP response (which might reveal more useful information.)
>
> Generally speaking, exceptions should capture as much information as
> possible, so this kind of enhancements are along this line and useful
> on its own.
>
> Between when a client application sends a request and receives a
> message, there are a lot of things that can go wrong. Having all those
> points report back diagnostic information via a general "response
> context" isn't actually that useful, because when you catch
> WebServiceException, the caller has no way of knowing what kind of
> details to look for.
>
>
>
> That said, if it's a requirement to keep the original behavior, then I
> need to do more thinking. I can't think of a clean way of doing this
> right now.
>
>